Mercurial > repos > guerler > springsuite

comparison spring_roc.py @ 31:3071750405c9 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit b306c715d8284c097220bc5c8461399fdc05eac7-dirty"
author guerler
date Tue, 24 Nov 2020 17:12:07 +0000
parents b0e195a47df7
children b300ddbbf9d0
comparison
equal deleted inserted replaced
30:b0e195a47df7 31:3071750405c9
208 prediction, _ = getReference(args.input, scoreCol=2) 208 prediction, _ = getReference(args.input, scoreCol=2)
209 209
210 # get subcellular locations from UniProt export 210 # get subcellular locations from UniProt export
211 locations = dict() 211 locations = dict()
212 if isfile(args.locations): 212 if isfile(args.locations):
213 regions = list()
214 if args.regions:
215 regions = args.regions.split(",")
213 with open(args.locations) as locFile: 216 with open(args.locations) as locFile:
214 for line in locFile: 217 for line in locFile:
215 searchKey = "SUBCELLULAR LOCATION" 218 searchKey = "SUBCELLULAR LOCATION"
216 searchPos = line.find(searchKey) 219 searchPos = line.find(searchKey)
217 if searchPos != -1: 220 if searchPos != -1:
218 uniId = line.split()[0] 221 uniId = line.split()[0]
219 locStart = searchPos + len(searchKey) + 1 222 locStart = searchPos + len(searchKey) + 1
220 locId = line[locStart:].split()[0] 223 locId = line[locStart:].split()[0]
221 if locId in ["Nucleus", "Membrane", "Cytoplasm"]: 224 if regions:
222 if uniId in filterA and uniId in filterB: 225 if locId not in regions:
223 locations[uniId] = locId 226 continue
227 if uniId in filterA or uniId in filterB:
228 locations[uniId] = locId
224 print("Found %d subcellular locations." % (len(list(locations.keys())))) 229 print("Found %d subcellular locations." % (len(list(locations.keys()))))
225 230
226 # estimate background noise 231 # estimate background noise
227 print("Estimating background noise...") 232 print("Estimating background noise...")
228 negative = set() 233 negative = set()
267 if __name__ == "__main__": 272 if __name__ == "__main__":
268 parser = argparse.ArgumentParser(description='Create ROC plot.') 273 parser = argparse.ArgumentParser(description='Create ROC plot.')
269 parser.add_argument('-i', '--input', help='Input prediction file.', required=True) 274 parser.add_argument('-i', '--input', help='Input prediction file.', required=True)
270 parser.add_argument('-b', '--biogrid', help='BioGRID interaction database file', required=True) 275 parser.add_argument('-b', '--biogrid', help='BioGRID interaction database file', required=True)
271 parser.add_argument('-l', '--locations', help='UniProt export table with subcellular locations', required=False) 276 parser.add_argument('-l', '--locations', help='UniProt export table with subcellular locations', required=False)
277 parser.add_argument('-r', '--regions', help='Comma-separated regions', required=False)
272 parser.add_argument('-e', '--experiment', help='Type (physical/genetic)', default="", required=False) 278 parser.add_argument('-e', '--experiment', help='Type (physical/genetic)', default="", required=False)
273 parser.add_argument('-t', '--throughput', help='Throughput (low/high)', default="", required=False) 279 parser.add_argument('-t', '--throughput', help='Throughput (low/high)', default="", required=False)
274 parser.add_argument('-m', '--method', help='Method e.g. Two-hybrid', default="", required=False) 280 parser.add_argument('-m', '--method', help='Method e.g. Two-hybrid', default="", required=False)
275 parser.add_argument('-o', '--output', help='Output (png)', required=True) 281 parser.add_argument('-o', '--output', help='Output (png)', required=True)
276 args = parser.parse_args() 282 args = parser.parse_args()